Tail-Hedging Cheat-Sheet
When (and what) to hedge — one table across 7 assets
July 20, 2026 — Cover / hub for the tail_hedge/ study
The whole
tail_hedge/study reduces to one decision: compare the VRP you pay to the asset’s break-even VRP. This page is the scannable summary; the S&P, VLCC and Sectors + quality names reports have the full work. Education/analysis, not investment advice.
The master table
| Asset | Ann. vol | maxDD | Long-run CAGR | Hold? | Break-even VRP | Live paid VRP | Hedge verdict |
|---|---|---|---|---|---|---|---|
| S&P 500 | 17% | −57% | +8.4% | ✅ best (diversified) | ≈ 0% | — (index, cheapest) | ❌ Hold / diversify |
| XLF Financials | 29% | −83% | +5.7% | ✅ | ≈ 0% | LEAPS puts thin | ❌ Hold |
| JPM (quality) | 38% | −74% | +10.3% | ✅✅ | ≈ 0% | ≈ 107% | ❌❌ Hold (puts far too dear) |
| AXP (quality) | 36% | −84% | +10.9% | ✅✅ | ≈ 0% | ≈ 99% | ❌❌ Hold (puts far too dear) |
| XLK Technology | 26% | −82% | +9.2% | ✅ | ≈ 27% | ≈ 24% | 🟡 Tactical — hedge only when crash risk is elevated (AI-bubble) |
| DHT VLCC | 48% | −97% | −6.2% | ❌ cyclical | ≈ 67% | ≈ 33% | 🟡 Only near a cycle TOP (CRule 1/5) |
| FRO VLCC | 61% | −98% | −4.7% | ❌ cyclical | ≈ 0% | ≈ 26–31% | ❌ Trim / FFA, not puts |
Break-even VRP = the max option over-pricing (IV/realized − 1) at which hedging still raises CAGR. Paid VRP = live option IV / trailing realized − 1 (Jul 2026 snapshot). All break-even numbers are crash-regime-dependent — the positive ones (XLK 27%, DHT 67%) come almost entirely from the 2000/2008 (XLK) and 2008–12 (DHT) crashes and fall to ~0% in calmer sub-windows.
The one decision rule
Hedge only if BOTH hold:
paid VRP < the asset's CAGR break-even VRP(with a margin for frictions), AND- you have a regime reason — a holdable asset in an elevated-crash-risk regime (Tech / AI-bubble), or a cyclical near its top (VLCC).
For 5 of the 7 assets both fail → HOLD, don’t hedge. Only XLK (tactically) and top-of-cycle DHT clear the bar today (XLK paid 24% < 27%; DHT paid 33% < 67% if you believe a big downturn is ahead).
Why: “deep AND frequent relative to drift”
What is “drift μ”? Drift is an asset’s deterministic upward trend over time — the directional part of the return once you strip out the random wobble. In the standard model
dS/S = μ·dt + σ·dW, μ is the drift (the trend) and σ is the volatility (the noise). Picture someone walking randomly on an escalator: their side-to-side sway is σ, but the escalator’s speed (μ) decides where they end up. Practically, the “Long-run CAGR” column above is the realized drift (CAGR ≈ μ − ½σ²): JPM/AXP/S&P have strong positive drift (a fast up-escalator you should just ride); DHT/FRO have negative drift (a down-escalator — holding only loses, so you must time it).
Tail-hedging is a race between premium bled while waiting (grows with the asset’s drift and its vol × VRP) and payoff harvested in crashes (grows with crash depth × frequency). The hedge only pays when the harvest beats the bleed:
- High drift raises the bar twice — it is the CAGR you must beat, and it pushes the underlying up and away from the strike so rolled puts expire worthless more often. That is why quality compounders (JPM/AXP, +10%) and the S&P are the worst hedge candidates despite deep 2008 tails.
- The same crash frequency is “enough” for a no-drift asset (VLCC) but “not enough” for a high-drift one (JPM). VLCC’s break-even is 67% because there is no drift to bleed against; but that also means there is no reason to hold it — so hedging it is pure cycle timing.
- Only Technology clears the bar as a hold-and-hedge asset — recurring deep crashes (2000, 2008, 2022) make the harvest frequent enough relative to its drift — and even that is regime-conditional (0% post-2010).
Net: convex tail-hedging is not a portfolio pillar; it is a conditional, regime-timed tool. For everything you’d actually want to hold, the drift is your friend and the cheapest “hedge” is diversification + time. Options tail-hedging earns its keep in only two places: Tech when you expect a drawdown, and a cyclical (VLCC) near its top — and if you must hedge a single quality name, hedge the index, not the name (single-name VRP ~100% vs index ~24%).
The full study
- S&P — 50-year backtest — the theory, passive-vs-ladder, why price (VRP) is destiny.
- VLCC (DHT/FRO) — win-rate-vs-VRP framework, break-even VRP, live options, cycle-position decision.
- Sectors + quality names (XLF/XLK, JPM/AXP) — hold-vs-hedge suitability, the drift-vs-hedge tension.
- AI-bubble report §11 — the crash-risk view that would justify a tactical XLK hedge.
Cross-asset cheat-sheet. Bilingual mirror: 中文版 →. Data + code: tail_hedge/. Education/analysis only — not investment advice.